Hypnotic suggestion and cognitiveneuroscience

David A Oakley

and Peter W. Halligan

Division of Psychology and Language Sciences, University College London, Gower Street, London, WC1E 6BT, UK

School of Psychology, Cardiff University, Tower Building, Park Place, Cardiff, CF10 3AT, UK

Thegrowingacceptanceofconsciousnessasalegitimatefield of enquiry and theavailability of functionalimaginghasrekindledresearchinterestintheuseofhypnosisandsuggestion to manipulate subjective experience and togain insights into healthy and pathological cognitivefunctioning.Currentresearchformstwostrands.Thefirstcomprises studies exploring the cognitive and neuralnature of hypnosis itself. The second employs hypnosisto explore known psychological processes using specifi-cally targeted suggestions. An extension of this secondapproach involves using hypnotic suggestion to createclinically informed analogues of established structuraland functional neuropsychological disorders. With func-tional imaging, this type of experimental neuropsycho-pathology offers a productive means of investigatingbrainactivityinvolvedinmanysymptom-baseddisordersand their related phenomenology.Introduction

Over the past decade, research involving hypnosis hasmade an important and unique contribution to the reﬁne-mentanddevelopmentofcognitiveneuroscience[1

–

3].Theadvent of sophisticated functional neuroimaging inparticular has made it possible to localize task-related,regionally speciﬁc brain activity associated with hypnosisas a mental state and the combination of hypnosis withsuggestion (hypnotic suggestion)[4,5]. Here, we outlinerelevant ﬁndings and review recent evidence of how hyp-nosis and suggestion provide a powerful tool for exploring normal and pathological psychological processes and con-ditions.

Hypnosis and suggestion

It is helpful ﬁrst to draw a distinction between ‘hypnosis’and the effects of suggestion. Operationally, ‘hypnosis’refers to a change in baseline mental activity after aninduction procedure and typically experienced at the sub- jectivelevelasanincreaseinabsorption,focusedattention,disattention to extraneous stimuli and a reduction inspontaneous thought[5]. Hypnotic induction procedurescomprise a set of verbal instructions that facilitate thisparticular mental state. Typical ‘hypnotic’ phenomena,such as alterations in sensory experience and motor con-trol,amnesiaandtheadoptionoffalsebeliefsabouttheself andtheenvironment,requirespeciﬁcsuggestions.Thereisgood evidence, however, that subjects can respond to sug-gestions of this sort without the need to employ formalinduction procedures. Indeed, the best predictor of thesuggestibility of an individual in hypnosis is their respon-siveness to the same suggestions outside hypnosis[6].Nevertheless, hypnotic induction procedures can increaseresponsiveness to suggestion, particularly if expectancy has been raised by explicitly labelling the procedure ‘hyp-nosis’[7]. Also, the effect of hypnotic suggestion can bemore evident at the level of brain activation. Derbyshire

etal.

[8], for example, showed that the same suggestions toincrease or decrease ﬁbromyalgia pain using functionalmagnetic resonance imaging (fMRI) produced greaterchanges in activation in pain-related brain areas whenparticipants were hypnotised compared to when they werenot, despite much less marked difference in reported subjective pain modulation between the two conditions. Forthe aforementioned reasons, most studies reported hereinvolve highly hypnotizable individuals, selected on thebasis of one or more of the standardised scales of hypnoticsusceptibility [9], and employ a formal hypnotic inductionprocedure. A second important distinction within hypnosisresearch[10]concerns studies attempting to elucidatethe neuro-cognitive nature of hypnosis (intrinsic studies)and those that use hypnotic suggestion as a means (instru-mental studies) of exploring a range of psychologicalphenomena (such as memory, perception, pain, hallucina-tion or voluntary control of action), including more recentattempts to develop hypnotic analogues for neurologicaland psychiatric conditions ([10

–

12].

Functional anatomy of hypnosis (

Intrinsic research

)

Although there have been some notable attempts to drawtogether neuropsychological and phenomenological evi-dence for hypnosis as an ‘altered state of consciousness’,with particular emphasis on the roles of anterior cingulateand dorsolateral frontal cortical areas[13], relatively littleis known from a cognitive neuroscience perspective aboutthe underlying processes involved in hypnotic experiencein the absence of suggestion

–

so-called ‘neutral hypnosis’.Usingpositronemissiontomography(PET),Rainville

etal.

[14]compared a no-hypnosis baseline condition with ahypnosis condition that produced a co-ordinated patternof activity involving brainstem, thalamus, anterior cingu-late cortex, right inferior frontal gyrus and right inferiorparietal lobule. These activations were interpreted asevidence that mental absorption (measured indepen-dently) is an experiential correlate of executive attentionalnetworks and central to the production of the hypnotic

Review

Corresponding author:

Oakley, D.A. (d.oakley@ucl.ac.uk ).

TICS-773; No of Pages 7

1364-6613/$

–

see front matter

experience. Similarly, a recent study (Q. Deeley

, et al.

unpublished) specifically explored the ‘default mode’ (thenormal resting state) of brain activation during hypnosisand found a different pattern of brain activity with adecrease in spontaneous conceptual thought compared tothe no-hypnosis conditions[4,5]. This study used hypnosisas a means of systematically modulating sustained atten-tion and stimulus-independent thought (a non-defaultactivity)withoutengagingaspecificgoal-directedcognitiveactivity under low demand stimulus conditions. Althoughfurther research is clearly needed, both preliminary sets of findingsareindicativeofadistinct‘defaultmode’ orneuralsignatureassociatedwithhypnosistogetherwithincreasesin mental absorption and reduction in spontaneous con-ceptualthought commonly reportedby hypnotised individ-uals. In a similar way, Fingelkurts

et al.

[15], using EEGmeasurements with a hypnotic virtuoso subject, foundalterations in local and remote functional connectivity between brain areas during neutral hypnosis, which werereplicated in the same subject one year later. Thesechanges imply a distinct hypnotic ‘state’ in which normalpatterns of communication between separate cognitivesystems are perturbed. In particular, they identified theweakening of remote functional connections as a possiblecorrelate for some of the unusual experiences that subjectscommonly report in hypnosis such as timelessness anddetachment from self. Converging evidence comes fromthe finding that administration of a hypnotic inductionprocedure is associated with spontaneous increases inerrors in a word and colour conflict test (the Stroop effect)in highly hypnotizable individuals[16]. The associatedincrease in activity in anterior cingulate cortex in theabsence of compensatory changes in left frontal corticalareas has been interpreted as evidence that hypnosis actsto decouple the normal relationship between conflictmonitoring and cognitive control[16].Collectively, these studies raise not only the possibility ofidentifyingdistinctpatternsofbrainactivationsattribu-table to hypnosis (including anterior cingulate cortex andfrontal cortical areas) but also indicate that these patternscomprise familiar components that can be found in many other cognitive tasks. In other words, a principled under-standing of hypnosis is possible without the implication of arcane or esoteric processes that has arguably slowed theuptake of ‘hypnosis’ as a cognitive tool for illuminating interesting scientific questions about memory, perception,attention and volition.

Functional modulation of cognitive performance withhypnotic suggestion (instrumental research)

Hypnosis and suggestionoffer a directmeans of studyingawide range of cognitive processes. One compelling line of research exploring attention and attentional conflict hasinvolved the suppression of the Stroop effect using a hyp-notic suggestion that disrupted lexical processing [17,18]and the elimination of the flanker compatibility effect by hypnotic suggestions that increased focal attention[19]. Inthe case of the Stroop studies, the suppression of atten-tional conflict was associated with reduced activity inanterior cingulate cortex and in visual cortical areaspossibly related to word recognition[20]. Congruent withchanges in visual cortical functioning seen in Stroop per-formance due to suggestion, Kosslyn

et al.

[21]reportedmodulationofactivityin colourprocessing(fusiform) areasof visual cortex after direct hypnotic suggestions to per-ceive a coloured stimulus as grey-scale or to experiencegrey-scale stimuli in colour. Also related to colour percep-tion, a recent study reporting the successful creation of grapheme-coloursynaesthesiainnon-synaesthetesbyhyp-notic suggestion[22]used the outcome to support the viewthat the unusual cross-modal interactions found insynaesthesia could be the product of disinhibition betweenbrain areas rather than hyperconnectivity.The Stroop studies relied in large part on the effective-ness of hypnotic suggestion to produce alterations in long established automatic linguistic processing. There is alsoconsistent, albeit less formal, evidence that hypnotic sug-gestion can affect performance relating to the use of numerical information. Theclassic ‘missingnumber seven’phenomenon, for instance, where it is suggested that thenumber seven ‘no longer exists’, produces disturbances tothe processes of counting and calculation. Although we areunaware of systematic studies, there are clear opportu-nities for using hypnotic suggestion not only to create, andreverse, tailored forms of dyscalculia as a way of exploring models of number processing but also to investigate othercognitive developmental conditions such as dysgraphia,dyslexia and dyspraxia.By contrast, there is a long history of memory researchbasedonthereversiblepost-hypnoticmemorylossthatcanbe produced by suggestion[23]. More recently, the focushas been on the selective and reversible loss of autobio-graphicalmemoryandthepreservationofimplicitmemory in post-hypnotic amnesia for autobiographical events[24

–

26], differences between directed forgetting and post-hyp-notic amnesia[27]and the creation of a robust and repea-tablelaboratoryanaloguefortheexperienceofde´ ja`vu[28].In a similar way, hypnosis has been employed as means of exploring ironic processes in thought suppression[29]. Inparticular, it has been found that hypnotised individualsdemonstrate a greater capacity for thought suppressionanddo not showthe usualparadoxical after-effects such asrebound, which has been interpreted as reﬂecting anincreased ability in hypnosis to manage cognitive load[23,30]. A recent neuroimaging study by Mendelsohn

et al.

[31], using a well established, reversible hypnoticsuggestion to suppress episodic memories (posthypnoticamnesia), reported alterations in brain areas responsiblefor long-term retrieval (occipital, temporal and prefrontalcortex).Thefindingsintheposthypnoticamnesiaconditionwere seen as evidence for the active inhibition of retrievalduring pre-retrieval monitoring.Studies with particular relevance to trauma-relatedmemory processes have incorporated hypnotic suggestionsto create analogues of emotional numbing and peri-trau-matic dissociation. Hypnotically induced emotional numb-ing [32]has been used to suppress affective responses toemotion-laden autobiographical memories without influ-encing the ability to recall the memories themselves[33],all of which supports the relative independence of the twoprocesses. Exploring the relationship between peri-trau-matic dissociation and flashback memories, Holmes

et al.

Review

Trends in Cognitive Sciences

Vol.xxx No.x

TICS-773; No of Pages 7

[34]used hypnotic suggestion to produce increased subjective dissociation during viewing ofa traumatic film. Thetheoretically predicted increase in memory flashbacks wasnotfoundand,whentakenwithotherconvergingevidence,challenges the assumption of a simple link between theexperience of dissociation at the time of trauma and sub-sequent proneness to flashback experiences. Adefiningfeatureofmosthypnoticallysuggestedeffectsis the report of subjective involuntariness[35]. This istypically demonstrated when hypnotic suggestion is usedto create ideo-motor responses; in which the hypnotisedperson’s own movements are experienced as occurring ‘allby themselves’ (anarchic control) or as the result of someexternal agency (alien control). In a recent study, Haggard

et al.

[36]used hypnotically suggested involuntary ﬁngermovement to investigate the relationship between theexperience ofvolition andthe subjective estimation oftimeof movement. The participants reported their hypnotically suggested involuntary movement and a comparable pass-ive movement as occurring closer in time to the actualmovement than for voluntary ﬁnger movements. Thisindicates that the normally experienced anticipatory underestimation of timing for voluntary movements doesnotdepend onthecentral motor preparatory phase,as thisis retained in ideomotor movement, but on the subjectiveawareness of an impending movement that underlies theexperience of intentionality, which is removed by the ideo-motor suggestion. In a related experiment, Blakemore

et al.

[37]investigated the brain processes underlying feelings of passivity or alien control produced by the hyp-notic suggestion that the participant’s hand was being moved up and down by means of a pulley. Subjectiveratings of involuntariness for the suggested passive move-ment and a comparable truly passive movement, in whichthe hand was actually moved by a pulley, were identical.Both the voluntary movements and the hypnotically suggested passive movements were associated withactivityinbrainareasassociatedwithleft-sidedmovementproduction, reﬂecting the fact that they were actively produced movements. The suggested passive movements,however, were also accompanied by greater activation inbilateral cerebellum and parietal cortex

–

also seen in thetruly passive movement condition and associated withfeelings of passivity and external agency. The findingsare consistent with the ‘forward model’ account of motorcontrol[37]and indicate that the hypnotic suggestion of passivity either prevented the generation of an effectiveforward model for the hand movement or interfered withthenormalsuppressionoffeedbackinformationassociatedwith voluntary action.Hypnotic analgesia suggestions are effective in alleviat-ing both clinical and experimental pain[38]and can selec-tively influence different aspects of pain experience. Whensuggestions were given in hypnosis to increase or decreasethe affective component of pain produced by an unvarying heat stimulus without affecting its perceived intensity, thesubjectively reported changes were accompanied by corre-sponding changes in activity in anterior cingulate cortex,whereas activation in other areas of the pain matrix wereunaffected[39]. A follow-up study [40]using the same experimental procedure showed the converse effect withactivity in primary somatosensory areas varying selec-tively in parallel with suggestions of changing pain inten-sity. Similarly, it is possible to demonstrate selectiveeffects of hypnotic suggestion on components of phantomlimb pain[41]. Suggested movements of the normally cramped and immobile fingers of a phantom arm[42]havebeenshownnotonlytoproducethesubjectiveexperienceof movement, with accompanying activation of contralateralsomatosensory cortex, but also with a reduction in theassociated pain. In a related study [43]in which thesuggested movements of the phantom limb were to com-fortable or uncomfortable positions, the movement wasagain accompanied by activity in contralateral somatosen-sory cortex and where the position was painful there wasactivation of pain related areas (thalamus, anterior cingu-late, posterior cingulate and lateral prefrontal cortex).These observations clearly provide converging evidencefor cognitive models that emphasize the contribution of different neurocognitive pathways in the affective andsensory components of pain, in addition to the role of feedback and control in the experience of pain. Again, the evidence emerging from these studies is thathypnosis (in this case combined with suggestion) producesits effects through controllable functional changes withinneurocognitive systems normally involved in mediating the psychological processes under investigation. In thisway, hypnotic suggestion also influences aspects of ‘phe-nomenological awareness’, which in turn can provideinsights about the structure and biological basis of normalcognitive processes. Recent studies have also exploited thepotentialthathypnoticsuggestionoffersforunderstanding clinical conditions.

Experimental neuropsychopathology (instrumentalresearch)

According to Zvolensky

et al.

[44], the term ‘experimentalpsychopathology’ was first used in the 1950s to describelaboratorybasedattemptstodevelopviableclinicalmodelsofbehaviouraldysfunction.Experimentalpsychopathology attempts to elucidate the cognitive processes that contrib-ute, either in whole or in part, to the aetiology, exacer-bation or maintenance of abnormal behaviour[44]. Although not intended to produce exact replications of thepsychologicaldisorder,theapproachfocusedonasmallnumber of key symptoms with the aim of identifying coreimpairments,understoodintermsof(functional)deficitstoknown normal cognitive processes. Experimental neurop-sychopathology can be seen as productively extending thiswellestablishedapproachtoincludeconditionsfrommain-stream cognitive neuropsychology [45].Liketranscranialmagneticstimulation(TMS),theuseof hypnoticsuggestionasanexperimentaltoolprovidesforthecreation of clinically informed analogues (virtual patientsrather than virtual lesions) delivered through intact cogni-tive neural systems, and enables cognitive neuroscientiststoselectivelymanipulatecomponentsofknowninformationprocessing processes with a view to assessing their impactoncognitiveoutput.Ifdisordersofpsychologicalfunctionaredefinedbyreferenceto‘normal’information-processingsys-tems[46], then credible reversible psychological disturb-ances produced in hypnosis (hypnotic analogues) can be

Review

Trends in Cognitive Sciences

Vol.xxx No.x

TICS-773; No of Pages 7